As a user, I'm happier with copyleft. I like to take my Marshall smart speaker as an example: that thing doesn't get any updates, ever. But it connects to the Internet. The app absolutely sucks, the connectivity is passable at best (often frustrating), but the hardware itself is nice (it looks nice in my living room and the sound is good when it works).
If all the open source software running inside that thing was GPLv3, Marshall would have to provide me with a way to patch it. So at the very least I could make security updates myself. But because Marshall used permissively-licenced dependencies, they locked it down in such a way that I can't do that.
The permissive licence helped Marshall, but for me as a user, the code may as well be proprietary.
If I use a copyleft licence (I like EUPL or MPLv2), it doesn't mean that they will open clean PRs, but at least they have to publish their changes in their own fork. It has happened to me that I could go read a fork, find a few things that were interesting and bring them back to my project.
With permissive licences, the risk is that those (typically businesses) who keep their fork open source probably don't see a lot of value in their fork, otherwise they would have made it private, "just in case".
The main difference is that GPL3 is a copyleft license, whereas MIT is not. Meaning that legally there is nothing in the license preventing a company from taking your code and using it for their purposes without having to contribute to improve the code.
the reason i picked mit is because rss is in a rough spot right now. the tech isn't mainstream, and big companies are trying to squash it since it doesn't drive engagement like the infinite scroll. anything that helps rss move forward is a win, and the mit license makes that easier.
> the reason i picked mit is because rss is in a rough spot right now.
I don't think another client is the solution, just saying. There's about three billion of them out there (though I don't dispute that yours might have something unique).